Changeset: 34479372
Quality Assurance on trees (type=broad_leafed => leaf_type=broadleaved)
Closed by LeTopographeFou
Tags
created_by | JOSM/1.5 (8800 fr) |
---|
Discussion
-
Comment from aseerel4c26
Hi LeTopographeFou, could you please comment on the relation of your changeset with https://wiki.openstreetmap.org/wiki/Automated_Edits_code_of_conduct ?
-
Comment from gormo
...especially on "Document and discuss your plans". Where is the documentation? Was there a discussion?
-
Comment from aseerel4c26
umm, there are much more similar changesets by LeTopographeFou. Where is the discussion for those changes?
-
Comment from aseerel4c26
... I've wrote a notification mail.
-
Comment from aseerel4c26
... and the DWG is aware too
-
Comment from LeTopographeFou
Thanks for your care aseerel4c26. Because the type=broad_leaved has been declared as deprecated on the Wiki, on TagInfo, and in JOSM (where I've had more than once warnings on this fact), I did not feel the need to open any discussion with the community neither to create any page to document a fact already approved somewhere else. Sorry if I thought wrong. If the DWG thinks that it is vandalism and/or if this property is not deprecated then I don't mind about reversing it and let other guys fixing it at their own pace.
-
Comment from fkv
The automated edits code of conduct does not depend on whether a tag is "deprecated". So-called deprecated tags are not necessarily wrong. First of all, many deprecations are done by single users who edit the wiki without prior discussion. Even if a formal proposal has been accepted by voting, the voters are just a handful of users, while the objects in the database have been created by thousands of users. Secondly, when you want to replace "old" tags by "new" tags, you need to make sure that no information is lost! With leave_type=broadleaved you lose the information whether the tree is a palm, so don't do that unless you add other tags like species=*. If you had discussed your edits beforehand, as you were obliged to, you would have been aware of that, and you certainly would not have done these mass edits at all.
It's also bad manners to change objects which were created by others without asking them.
I was already preparing a revert of one of your changesets (that damaged objects that I had carefully created) when I noticed that a discussion is ongoing here. Given that the DWG was informed 10 months ago, why are the abusive changesets not reverted yet? What are they waiting for?
-
Comment from SomeoneElse
@fkv When I looked at the changes in here back in October last year I didn't see any evidence of data loss (though I only looked at a sample of the 5341 nodes).
Whilst a change like type=broad_leafed => leaf_type=broadleaved should have been discussed beforehand, in case someone was aware of an edge case where the change could be wrong, in retrospect I can't think of one.
Can you give me an example of where something was removed that should not have been? -
Comment from fkv
Just take the first few nodes in this changeset as examples. They are located on Plaza de la Ciudad de la Habana, where there are palms and other trees. The non-palms were changed from type=broad_leafed (old spelling variant, but that's another issue) to leaf_type=broadleaved, losing the information that they are not palms. Interestingly, the palms were left untouched, they are still tagged with type=palm. I guess that LeTopographeFou didn't know what to do with them, or his validator didn't blink. Given that there are still 18275 type=palm and 57508 type=broad_leaved instances in the database (according to Taginfo; even after all those validator-triggered mass edits, not only by LeTopographeFou), it's obvious that these tags are still widely used to distinguish palms from other trees, and that the wiki is wrong in declaring them deprecated.
-
Comment from LeTopographeFou
Hi, this changeset was made a long time ago and I thought it has been reverted (with some others of the same type and of the same period). I'm going to ask for a revert by mail if it is still feasible.
-
Comment from SomeoneElse
@fkv thanks; I wasn't aware of the use of "type" as meaning "not a palm tree".
@LeTopographeFou thanks; I've got the message requesting a revert. Before doing it I'll ask around to see if there's any better way to indicate "not a palm tree".
As I understand it (from reading wikipedia just now) Palm Trees https://en.wikipedia.org/wiki/Arecaceae are Angiosperms , as are other broadleaved trees such as https://en.wikipedia.org/wiki/Oak . However Palm trees are in https://en.wikipedia.org/wiki/Monocotyledon and typical temperate broadleaved trees in https://en.wikipedia.org/wiki/Eudicots . For comparison conifers are in https://en.wikipedia.org/wiki/Gymnosperm . Strictly speaking therefore "leaf_type=broadleaved" would be correct for both palms and non-palms, but were all of the examples in this changeset definitely NOT palms? That's what I'll need to investigate and the tagging that I'll ask about, and I'd like to ask before the revert so that if there are other changes to be made they can be done at the same time. -
Comment from fkv
I was asking for a revert, but it might be better to restore the type=* tags while keeping the leaf_type=* tags. It does not hurt to have both.
You could use taxon=* to mark palms, but there is no single taxon that includes all non-palm broad-leaved trees. Eudicots would be missing magnoliids, Austrobaileyales etc. You also need to consider that these classifications are permanently disputed and changing. The lowercase, suffixless taxon names like "eudicots" do not even indicate a taxonomic level. It's impossible for average mappers as well as data users to use such tags. -
Comment from SK53
I tend to agree with @fkv on all issues: 1) large scale updates often have unpredictable side effects; 2) originally we distinguished between conifers/dicots aka broad-leaved and palms (and possibly one or two others) using a range of tags wood/type etc. Ultimately these have been replaced by leaf_type. However, the palm issue (and also Gingko) highlight that leaf_type=broadleaved doesn't quite do what I think most of us expected. As @fkv nicely shows modern systematics doesn't help either. Whereas I would expect ordinary mappers to not know that Gingkos are gymnosperms; I think most should be able to distinguish palms.
To summarise: the tagging issues involved are rather more complex that initially apparent. Reversion at least would retain mapped information. Further discussion is needed on general values for leaf_type before such wholesale changes are introduced automatically.
-
Comment from SomeoneElse
OK, done: See https://www.openstreetmap.org/changeset/41192016 .
-
Comment from LeTopographeFou
@SomeoneElse: Thank you!
- Albvereinslinde 1964 (1387847717), v2
- Alte Buche (1207186319), v2
- Alte Eiche (684150669), v4
- Alte Linde (698448650), v5
- Andres Eichenauer Buche (746857838), v3
- Arbre de la Liberté (1390691274), v4
- Auferstehungslinde - Naturdenkmal (846271539), v7
- Big Banyan Tree (1325967550), v2
- Bildeiche (448380841), v5
- Birnbaum am Binsenrasen (1335988250), v3
- Birnbaum am Lerchenberg (638865135), v2
- Birnbaum zwischen den Dörfern (1128850878), v2
- Bismarkeiche (1112204971), v3
- Bligger-Linde (861612695), v7
- Buk Anton (475704144), v4
- Chêne rouge (618588624), v6
- Chêne rouge (621774435), v2
- Cunnersdorfer Linde (462077231), v5
- Deutschland-Linde (937541294), v3
- Dick Buche (1707900597), v3
Welcome to OpenStreetMap!
OpenStreetMap is a map of the world, created by people like you and free to use under an open license.
Hosting is supported by Fastly, OSMF corporate members, and other partners.
https://openstreetmap.org/copyright | https://openstreetmap.org |
Copyright OpenStreetMap and contributors, under an open license |